Work out work sampling study size instantly with clear inputs, formula shown and shareable results.
Work sampling estimates a proportion, so sample size follows the binomial formula n = z^2 x p x (1-p) / e^2, where e is the absolute accuracy in proportion points. Studies of rare activities need far more observations, because the required sample rises steeply as p falls for a fixed relative accuracy.
Work sampling size
n = z^2 x p x (1 - p) / e^2
Duration
Study days = observations required / observations per day
Any regular pattern lets operators anticipate the observer and lets cyclical work bias the result. Draw observation times from a random schedule each day.
Use a pilot study of 100-200 observations, or the most pessimistic plausible value. Using p = 0.5 gives the largest and therefore safest sample size.